On the Reliability of Cascade Size as a Virality Measure

Online social networks have become a major news source as well as a platform for information dissemination. The spread of information on social networks has a recognizable impact on political and marketing campaigns and emergency situations. However, the credibility of the information propagating through social networks is in threat because of social spammers, misinformation, and crowdturfing. In this paper, we propose an approach to judge the reliability of information cascade size in social networks, as a measure of virality, by observing the network k-core sizes. The proposed approach is based on previous research that proves the best information spreaders are consistently located in the k-core. The proposed approach was examined on randomly generated graphs and on four real life Twitter datasets. To further examine the approach, an incremental number of fake tweets were injected into two of the datasets, and the effect was observed. Results show that the proposed approach has the ability to estimate the trustworthiness of an observed meme size cascade, and hence signify the real impact of a meme without the misleading effect of spammers and crowdturfers on the cascade size.